- Pharo-users - lists.pharo.org

Re: [Pharo-users] New book for Pharo :)

by HernÃ¡n Morales Durand

2015-01-28 14:21 GMT-03:00 stepharo <stepharo(a)free.fr>: > Didier Besset offered his great book "Object-Oriented Implementation of > Numerical Methods > An Introduction with Smalltalk and Java" to the community. > > We would like to thank Didier Besset for his great book and for his gift > of the source and implementation to the community. > > This is indeed, great news. > > You can find > â¢ Archive of the original book, with code in both Java and Smalltalk > https://github.com/SquareBracketAssociates/ArchiveOONumericalMethods > > https://github.com/SquareBracketAssociates/ArchiveOONumericalMethods/ > blob/master/NumericalMethods/2015-Jan-WholeBookST-Java.pdf > > > â¢ An abridged version of Didierâs book, without the Java > implementation and reference; our goal is to make the book slimmer and > easier to read. The implementation presented in this book is part of the > SciSmalltalk library. > > The implementation presented in his book is part of the NumericalMethods library. You want to promote SciSmalltalk, then first respect others. NumericalMethods could be part of SciSmalltalk, but it is the library where code was originally released. Cheers, HernÃ¡n

Jan. 28, 2015

Re: [Pharo-users] Building 3D shapes in Roassal with interaction. Unexpected behaviour?

by Ronie Salgado

Hello Nicolas, Thanks for pointing this bug. I just fixed it in Woden-Core-RonieSalgado.64 . Can you try again? Greetings, Ronie 2015-01-28 10:40 GMT-03:00 Nicolas Lusa <nicolas.lusa(a)usi.ch>: > v := RWView new. > > e := (RWPyramid new) element. > e on: RWMouseButtonDown do: [ :ev | > ev element shape color: WDColor blue. > ev element changed. > ]. > v add: e. > > g := (RWCube new) element. > g on: RWMouseButtonDown do: [ :ev | > ev element shape color: WDColor green. > ev element changed. > ]. > g translateBy:(WDVector3 newX: 4.0 y: 0.0 z: 0.0). > v add: g . > > f := (RWCylinder new) element. > f on: RWMouseButtonDown do: [ :ev | > ev element shape color: WDColor red. > ev element changed. > ]. > f translateBy:(WDVector3 newX: 2.0 y: 0.0 z: 0.0). > v add: f. > > v addInteraction: RWMouseKeyControl. > v camera position: (WDVector3 x: 0.0 y: 0.0 z: 3.0). > v open. >

Jan. 28, 2015

Re: [Pharo-users] Why single inheritance?

by Cameron Sanders

Laura, Obviously, I should have let the experts answer, and I did not answer all of your questions in that last quick-off-the-cuff diss-the-strongly-typed-languages answer. My apologies. For me, i like to keep the namespace clean. And one must admit, with multiple inheritance, things can get confusing, and in some cases, the wrong methods come into play. If one breaks free of inheritance as a means to gain polymorphism, code re-usability can grow considerably, and the ability to substitute classes to be used by a 3rd party libraries grows with it. I generally prefer delegation(/decoration) approaches to inheritance, in part just to control the namespace and "interface" of a class. Traits are still not widely available in different smalltalks, are they? -Cam On Wed, Jan 28, 2015 at 10:41 AM, Laura Risani <laura.risani(a)gmail.com> wrote: > Hi all, > > What is the explanation why Smalltalk designers preferred single > inheritance+traits to multiple inheritance? Why is the former better than > the latter? > Do traits let you share state also or only behavior? > > Best, > Laura >

Jan. 28, 2015

Re: [Pharo-users] Why single inheritance?

by Cameron Sanders

Because smalltalk methods require so little code that overt-delegation is both easy to author and to understand. If on the other hand, one needs to repeatedly tell the compiler what types of data shall be used used, and then insert numerous delimiters, such delegation becomes onerous... and ... then one would beg for a compiler solution. -cam On Wed, Jan 28, 2015 at 10:41 AM, Laura Risani <laura.risani(a)gmail.com> wrote: > Hi all, > > What is the explanation why Smalltalk designers preferred single > inheritance+traits to multiple inheritance? Why is the former better than > the latter? > Do traits let you share state also or only behavior? > > Best, > Laura >

Jan. 28, 2015

Re: [Pharo-users] Why single inheritance?

by Damien Pollet

Mostly, multiple inheritance is difficult to implement well in a simple enough form. The only language I know of that has a reasonably usable version of it is Eiffel (probably Nit too, but I don't know enough about it http://nitlanguage.org) There is the diamond inheritance problem, the problem that inheritance (even in Smalltalk) mixes subtyping (interface extension) and reuse of implementation, problems of how to linearize method lookup, etc. On 28 January 2015 at 16:41, Laura Risani <laura.risani(a)gmail.com> wrote: > Hi all, > > What is the explanation why Smalltalk designers preferred single > inheritance+traits to multiple inheritance? Why is the former better than > the latter? > Do traits let you share state also or only behavior? > > Best, > Laura >

Jan. 28, 2015

Re: [Pharo-users] Supporting Start up around Pharo

by stepharo

This includes one year salary. But it requires building a case and being motivated :) Le 28/1/15 11:51, stepharo a Ã©crit : > Hi guys > > if you are considering to create a start up around Pharo, we could > find one year financial support > if this is serious and that you would like to come to work around > Lille or France. > > Stef > >

Jan. 28, 2015

New book for Pharo :)

by stepharo

Didier Besset offered his great book "Object-Oriented Implementation of Numerical Methods An Introduction with Smalltalk and Java" to the community. We would like to thank Didier Besset for his great book and for his gift of the source and implementation to the community. You can find â¢ Archive of the original book, with code in both Java and Smalltalk https://github.com/SquareBracketAssociates/ArchiveOONumericalMethods https://github.com/SquareBracketAssociates/ArchiveOONumericalMethods/blob/m… â¢ An abridged version of Didierâs book, without the Java implementation and reference; our goal is to make the book slimmer and easier to read. The implementation presented in this book is part of the SciSmalltalk library. https://github.com/SquareBracketAssociates/NumericalMethods https://github.com/SquareBracketAssociates/NumericalMethods/blob/master/201… â¢ SciSmalltalk library https://github.com/SergeStinckwich/SciSmalltalk. Both book versions are maintained by St Ìephane Ducasse and Serge Stinckwich. 27 Janvier 2015

Jan. 28, 2015

Re: [Pharo-users] Waiting object

by Ben Coman

On Wed, Jan 28, 2015 at 11:31 PM, Laura Risani <laura.risani(a)gmail.com> wrote: > Hi Ben , > #fork let me implement the overall functionality i wanted, but i had to > introduce some minor changes into the design to reflect the concurrent > nature of the msg. So i still feel curious about how one could implement > this msg > > Waiting>> wait : seconds > "i take the flow control during 'seconds' without halting image morphs and > only then return to the sender " > There is only one Morphic UI loop process which repetitively cycles for Morph stepping and the main interactive UI. Use "Tools > Process Browser" to take a look at that process (priority=40) and its stack and you'll WorldMorph>>doOnceCycle. If you are curious, you can put a "self haltOnce" in that method and invoke it using "System > Enable halt/inspect once" and it will spawn that fault in that cycle off to another process you can trace through, while the main process continues to cycle. However if you perform a non-concurrent #wait or busy calculation, then the UI loop is prevented from looping at its usual speed (see #interCyclePause). cheers -ben > > Best, > Laura > > On Tue, Jan 27, 2015 at 7:00 PM, Ben Coman <btc(a)openinworld.com> wrote: > >> Hi Laura, Can you provide some more context by pasting code for how you >> expect both your UI and calculation methods to interact? What is the >> application? >> >> cheers -ben >> >> On Wed, Jan 28, 2015 at 5:49 AM, Laura Risani <laura.risani(a)gmail.com> >> wrote: >> >>> Hi all, >>> Thank you for your answers. >>> >>> But i yet can't see how to implement the method i wanted, because #fork >>> schedules the process and immediately returns but i'd like to wait and only >>> then return the answer. >>> >>> >>> Best, >>> Laura >>> >>> On Tue, Jan 27, 2015 at 3:46 PM, Laura Risani <laura.risani(a)gmail.com> >>> wrote: >>> >>>> Hi all, >>>> >>>> I need an object/method like this one >>>> >>>> Waiting>> wait : seconds >>>> "i take the flow control during 'seconds' without halting image morphs >>>> stepping/handling (as Delay>>wait: does) and then return to the sender " >>>> >>>> How can i implement it? >>>> Where can i borrow for reuse an existing one from? >>>> >>>> >>>> Best, >>>> Laura >>>> >>> >>> >> >

Jan. 28, 2015

PostGIS in Pharo

by Esteban A. Maringolo

Is there anybody using PostGIS in Pharo? Regards! Esteban A. Maringolo

Jan. 28, 2015

Re: [Pharo-users] dirty packages

by Esteban Lorenzano

You can try rebuilding the rpackage structure. 1) close all browsers 2) execute RPackageOrganizer default initializeFromMC Esteban > On 28 Jan 2015, at 16:50, Usman Bhatti <usman.bhatti(a)gmail.com> wrote: > > Hello, > > I have two packages MyPackage and MyPackage-Ext. When I load MyPackge-Ext, MyPackage becomes dirty and do not have any clue why. When I try to see changes in MyPackage in Monticello Browser, the package becomes clean again. > > There are two possibilities why MyPackage-Ext makes MyPackage dirty: > -> There are names conflict of some sort (MyPackage-Ext can be seen as a category for MyPackage). > -> Overriding of the extension methods introduced on the third package. Now I wrote a small script to verify it: > firstCollection := (RPackage organizer packageNamed: 'MyPackage') extensionMethods values flatten. > secondCollection := (RPackage organizer packageNamed: 'MyPackage-Ext') extensionMethods values flatten. > (secondCollection intersection: firstCollection) > The above script gives empty result so there are no overriding selectors. > > Any ideas how can I hack into Monticello to see what is causing MyPackage to become dirty while I load the second one manually. May be by logging the override load definitions on Transcript? > > In general, will it be possible to see overriding without these disappearing when doing changes on a dirty package? > > > tx in advance. > > Usman > >

Jan. 28, 2015